Field
The present disclosure relates generally to a method for detecting a voice and an electronic device using the same.
Description of Related Art
A user of an electronic device can activate or trigger the electronic device that includes a portable terminal through voice detection (or voice wakeup). The electronic device can execute a specific function through detection of user's voice. In general, voice detection may be implemented in the form of keyword detection for detecting a specific word spoken by the user.
It is general that keyword detection is performed using a pattern matching technology that extracts features from a voice that is input to a microphone of an electronic device and matches the extracted features with a pre-generated model of a keyword to be detected. In this case, if an ambient noise or an echo on an output end is included in the voice that is input to the microphone, the keyword detection performance may deteriorate. In particular, in the case where a specific sound source (e.g., music, Text To Speech (TTS), or Automatic Response Service (ARS)) is reproduced in an electronic device that performs the voice detection, the specific sound source may barge into the microphone of the device (e.g., echo or the like) to cause malfunction of the keyword detection. For example, if a user speaks a specific keyword while an echo is generated, the keyword spoken by the user and the echo may be mixed with each other to cause the keyword detection performance to deteriorate.